Pages: [1]   Go Down

Author Topic: 20D noise  (Read 1598 times)

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
20D noise
« on: November 17, 2004, 04:20:35 pm »

Thiery,

A) The technical answer is that no digital camera has any true ISO speed: there is only an ISO standard defining film speed, based on things like shadow handling and contrast, and this is reported in both ASA Exposure Index units like 100, 125, 160, etc. and DIN degree units 21º, 22º, 23º, etc. (Note that films always report the speed as something like "ISO 100/21º", combining the ASA and DIN units. Despite what numerous photographic books say, it is not true that the ISO standard simply adopted the ASA scale: they merged ASA and DIN, for the sake of international bureaucratic harmony.)

There is no such ISO standard for the sensitivity of electronic sensors yet. (There is a "base ISO" standard for the overexposure limit of a sensor, but that is the opposite thing to a sensitivity measure.)


 The practical answer is that almost everyone (incorrectly?) refers to those ASA units for measuring Exposure Index by the name "ISO", regardless of whether a high "ISO" setting has far better or worse shadow handling than required of a film with that ISO speed rating.

My observation is that at minimum exposure index settings like 100 or 200, many or all DSLRs have distinctly better shadow handling than required by the ISO film standard, so that a real "ISO" speed would be higher. Also, different signal processing at different speed settings could maintain the required shadow handling over a wide range of speed settings, at the cost of resolution: in particular, Fuji seems to do this in their DSLRs.

I am even willing to go so far as to speculate than when Kodak and Dalsa  state "ISO" speed ranges for their sensors, sold mostly to scientific, medical and high end photographic companies, they are stating a range over which something like the ISO shadow handling standard is met. Kodak is fairly cautious with their stated speed ranges, like 100-800 for the sensor in the Olympus E-1, despite Olympus also offering noticably noisy 1600 and 3200 settings. Images from that sensor at up to EI 800 do seem to print with little visible noise, while EI 1600 and 3200 produce noticably noisy prints unless noise reduction processing is applied.

P. S. noise reduction processing can arguably give a genuine "ISO speed boost", at the usual cost in resolution.
Logged

jpoll

  • Newbie
  • *
  • Offline Offline
  • Posts: 6
20D noise
« Reply #1 on: November 18, 2004, 07:04:31 pm »

Although not about Canon sensors you might want to check out the following link concerning Kodak and ISO.

Kodak ISO Measurement
Logged

thierrylegros396

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1947
20D noise
« Reply #2 on: November 17, 2004, 12:02:43 pm »

If I'm right the 20D has not a 3200 ISO immediately accessible.

So is it a real 3200 ISO, or a simulated by changing the exposure ?!
Logged

thierrylegros396

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1947
20D noise
« Reply #3 on: November 18, 2004, 11:05:46 am »

Thanks for your quick answer

Thierry
Logged
Pages: [1]   Go Up